Are you interested in finding out whether a computer can tell from images of the brain whether the scanned person suffers from a mental disorder, and if yes, from which? Do you want to know whether combining such images with clinical information can predict, which therapeutic intervention may be the most effective and how a patient will feel a year from now?
These are the questions at the cutting edge of precision psychiatry, and if you are intrigued by them, this course may be for you.
This course will be taught in three parts:
Part I: self-paced machine learning crash course from google colab accompanied by weekly live-video tutorials via zoom.
Part II: More in depth analysis of medical images based on course material from NVIDIA's Deep Learning Institute with exercises in Amazon's cloud accompanied by weekly live-video tutorials via zoom or classroom-meetings.
Part III: Group projects on using functional and structural MRI scans for diagnosis and prognosis accompanied by weekly live-video tutorials via zoom or classroom-meetings.
Audience
Master and Doctoral Students from Medicine, Psychology, Physics, and Computer Science. Bachelor students, please contact me (jens.schwarzbach@ukr.de) before signing up.
Prerequisites: basic programming knowledge in Python, some basic knowledge in ML helpful.
Structure
Self-paced selfstudy of the material indicated for each session. The sessions themselves are interactive videoconferences (Thursdays at 2.15PM-3.45PM via zoom) in which we discuss problems with the concepts or materials, extensions, and do code reviews.
Links to zoom meetings will be announced an hour before the start. Please log in 10 minutes ahead of schedule.